7 Workflows and Extensions

7.1 Related Nomenclature Variables

	Same Measure	Diff Measure
Same Classification	1: Already harmonised Could be row-concatenated together Likely different observation units	3: Distinct but related variables Share xmap from_set, but not necessarily weights
Diff Classification	2: Ex-post harmonisation Related observational units (e.g same country, different years) Data must be transformed before concatenation	4: Distinct variables No harmonisation possible

7.2 Multiple Transformations

7.2.1 Multi-Step Transformation (WIP)

Consider the task of transforming some data from classification A to C, and imagine you only have panel maps for the steps A-to-B and B-to-C. It is straightforward to chain these transformations together.

7.2.1.1 Stylized Code

7.2.1.2 Functions

7.2.1.3 Tests

7.2.2 Multi-Group Transformations (WIP)

Consider the workflow where you want to apply the same transformation to multiple subsets of a larger dataframe. For example, trade data grouped by country.

7.2.3 Mutli-Map-Group Transformations (WIP)

Consider datasets where you might want to apply variations of the same map to distinct subsets of a larger dataframe.

7.3 Visualising Panel Map Representations

See viz-panel-maps for more:

7.3.1 Panel Maps as Graphs

pm_BA |> plt_pm_sigmoid(std_A, std_B, weight) +
  scale_fill_brewer(palette="RdPu", direction = -1)

7.4 Custom Mapping Weights

Mapping weights could be based on:

• Reference data – e.g. population share, past/future proportions
• Domain expertise
• ???

7.4.1 Manual Design and Encoding

7.4.2 Use Reference Proportions